Fish communities on the world's warmest reefs: what can they tell us Fish communities on the world's warmest reefs: what can they tell us about the effects of climate change in the future? about the effects of climate change in the future?

To examine the role of climatic extremes in structuring reef ﬁsh communities in the Arabian region, reef ﬁsh communities were visually surveyed at four sites within the southern Persian Gulf (also known as the Arabian Gulf and The Gulf), where sea-surface temperatures are extreme (range: 12–35 ◦ C annually), and these were compared with communities at four latitudinally similar sites in the biogeographically connected Gulf of Oman, where conditions are more moderate (range: 22–31 ◦ C annually). Although sites were relatively similar in the cover and composition of coral communities, substantial differences in the structure and composition of associated ﬁsh assemblages were apparent. Fish assemblages in the southern Persian Gulf held signiﬁcantly lower estimates of abundance, richness and biomass, with signiﬁcantly higher abundances of smaller sized individuals than Gulf of Oman assemblages. Functionally, southern Persian Gulf sites held signiﬁcantly lower abundances of nearly all the common ﬁsh trophic guilds found on Gulf of Oman sites, although higher abundances of herbivorous grazers were apparent. These results suggest the potential for substantial changes in the structure of reef-associated ﬁsh communities, independent of changes in habitat within an environment of increasing ﬂuctuations in oceanic climate.


INTRODUCTION
There is now a wealth of information showing that ecological communities will change substantially with increasing fluctuations in global climate, especially within tropical marine areas (Hoegh-Guldberg, 1999;Carpenter et al., 2008;Gilman et al., 2008).Fluctuations in a range of physical variables, mediated by changes in global climate (e.g.sea-surface temperature and salinity) are predicted to affect directly the abundance, diversity, composition and demographic structure of the dominant habitat-forming corals in tropical marine communities, which will have indirect repercussions for associated fishes as a result of changes in habitat availability and quality (Purkis & Riegl, 2005;Baker et al., 2008;Bradbury, 2008).There is increasing evidence, however, that fluctuations in physical variables associated with climate change may also have direct effects on a much wider array of tropical marine organisms than previously envisaged (Moore et al., 2008;Przeslawski et al., 2008), mediated through biochemical changes in their physiology, growth, reproduction and, ultimately, survival (Hurst, 2007;Chown & Gaston, 2008;Figueira et al., 2009).Recent research indicates that changes in physical variables associated with climate change can have direct effects on individual fish physiology through reductions in growth, reproduction and survivorship (Munday et al., 2008;Nilsson et al., 2009), with the implication that these individual changes will result in larger changes to the wider fish communities in tropical systems (Munday et al., 2008;Nilsson et al., 2009;Wilson et al., 2010).Given that fishes are considered among the most important fauna in tropical marine systems in terms of both economic and societal value to humans (Jackson et al., 2001), understanding the implications of climate change on fish communities is of paramount importance.
In order to understand the potential effects of fluctuating physical extremes on fish communities in the future, it is appropriate to examine communities in areas already exposed to physical oceanic extremes.The north-eastern Arabian Peninsula is bounded by the Persian Gulf (also known as the Arabian Gulf and The Gulf) and the Gulf of Oman, with each distinguished by different oceanographic conditions (Sheppard et al., 1992;Coles, 2003).The Persian Gulf is a shallow (mean depth <36 m), semi-enclosed basin connected to the Gulf of Oman and the wider Indian Ocean through the narrow Strait of Hormuz (Sheppard et al., 1992).As a result of its shallow depth and restricted water exchange, the Persian Gulf is characterized by extreme sea-surface temperatures (SSTs) ranging from 12 • C in winter to summer highs >36 • C (Sheppard et al., 1992;Sheppard, 1993;Riegl, 2001;Sheppard & Loughland, 2002).Hyper-saline conditions occur year-round with observations as high as 43 common (Reynolds, 1993), with moderate to low primary production found within the southern Persian Gulf [average of 1•0 mg m 3 chlorophyll a levels using remote sensing data collected by SeaWiFS (September 1997to May 2006) and MODIS-A (July 2002 to May 2006) satellite sensors] (Nezlin et al., 2007).In comparison, the adjacent Gulf of Oman has deeper waters that are well mixed with those of the Indian Ocean, resulting in relatively mild seasonal changes in SSTs ranging from summer highs of 32 • C to winter lows of 22 • C (Coles, 1997;Bohm et al., 1999;Schils & Wilson, 2006).Fluctuations in salinity are relatively mild (35-37), while substantial upwelling along the coastal margin leads to a relatively high mixed layer primary production of 900 mgC m −2 day −1 (Brock et al., 1994;Nezlin et al., 2007).These two systems make for an interesting and appropriate comparison for the effects of changing climate on oceanic processes, as they are biologically and geographically connected with one another, but experience such different environmental conditions.Additionally, given that the Persian Gulf reefs persist despite experiencing the highest variability in annual temperature encountered by any coral reefs (Kinsman, 1964;Sheppard et al., 2000), they may serve as a useful proxy for the implications of changing ocean temperatures on a global scale (Riegl, 2003).
Despite environmentally different regimes between the two regions (Coles, 1997(Coles, , 2003)), there are broad similarities in benthic community structure between the Persian Gulf and Gulf of Oman.The Persian Gulf forms a biogeographic subset of the Indian Ocean, with close faunal affinity with communities in the Gulf of Oman (Sheppard, 1987;Sheppard & Sheppard, 1991).In addition, coral reef communities found within the southern areas of the Persian Gulf (mainly encapsulating the United Arab Emirates), although showing lower diversity are highly similar topographically to those found at similar latitudes within the Gulf of Oman.Within both regions stony coral cover is limited to areas of hard substratum, which can result in relatively patchy coral distribution (Wilson, 2007;Burt et al., 2008).Within areas suitable for coral growth, substantial coral reef communities will develop in both regions, and can be dominated by a range of branching, columnar, encrusting, laminar and massive morphologies (Sheppard et al., 1992;Coles, 2003).
This research examined the structure of coral reef communities throughout the north-eastern Arabian Peninsula, encapsulating areas both within the southern Persian Gulf and at similar latitudes within the Gulf of Oman.In particular, the structure and composition of reef-associated fish communities was compared between regions.To avoid the confounding factors of benthos v. environmental factors, this study focused on comparing fish communities within areas of similar coral structure.It was hypothesized that within areas of similar coral reef structure substantial differences in the structure and composition of fish communities would be apparent.In line with previous assessments of biodiversity estimates between the Persian Gulf and Gulf of Oman, it was predicted that fish communities within the southern Persian Gulf were more likely to show reduced diversity than Gulf of Oman communities, due to the relatively recent geological origin of the Persian Basin as well as the background physiological stress associated with extremes in physical variables (Price, 2002;Coles, 2003;Schils & Wilson, 2006).It was also predicted, however, that physiological stress associated with the physical extremes of the Persian Gulf may also have substantial negative effects on species growth (Hurst, 2007), with concomitant negative effects on community biomass and size structure.It was also predicted that there would be little difference in the guild structure of communities between regions, with the majority of differences associated with site-specific differences in available food and habitat.

MATERIALS AND METHODS
Between October and November 2008, the structure and composition of coral reef communities was examined between the southern Persian Gulf (Ras Ghanada, Sadiyaat, Ras Hasyan and Saih Al-Shaib) and Gulf of Oman (Dibba, Al-Aqah, Al-Ghattan and Al-Jazira) (Fig. 1).Each site was similar in terms of depth (c.5-6 m), distance to shore (c. 100 m) and latitude (Fig. 1).Logistic difficulties precluded multiple surveys of coral reef communities in both regions throughout the year.In addition, substantial reductions in the abundance of Persian Gulf fish communities, due to seasonal migrations of fish communities off reefs, occur in the winter months (Coles & Tarr, 1990;Burt et al., 2009).Therefore, to reduce any seasonal bias in fish community abundance between regions, all surveys were conducted during the northern hemisphere's summer months.Fishes were counted visually by scuba along eight 5 m × 30 m belt transects.All fishes were identified to species (Randall, 1995) and categorized into 50 mm interval size classes (total length, L T ) for later conversion to biomass based on species-specific growth coefficients (Froese & Pauly, 2009).Fish species were also classified by primary feeding guild (i.e.corallivore, grazer, invertivore, piscivore, piscivore-invertivore, planktivores and scraper and excavator) (Randall, 1995;Froese & Pauly, 2009).Coral communities were sampled using 0•25 m 2 quadrats photographed at 3 m intervals on the same transects sampled for fishes (11 quadrats per transect) (Burt et al., 2009).Photographs were analysed using 50 random point intercepts in coral point count with Excel extensions (CPCe) software (Kohler & Gill, 2006), with all corals identified to species and classified by dominant morphology (Veron, 2000).Coral morphological categories were branching, columnar, encrusting, laminar and massive.
Prior to multivariate analysis, all fishes and coral species occurring in <5% of samples were removed in order to reduce the influence of outliers (Tabachnick & Fidell, 2001;McCune & Grace, 2002).To examine whether there were significant differences in coral cover between regions and among sites within regions, data were square-root transformed and compared using nested ANOVA.To compare the structure of fish communities between regions and among sites within regions, data were log 10 (n + 1) transformed and the abundance and biomass compared using nested ANOVA, with post hoc Tukey's tests used to identify significantly different groups.As estimates of abundance and richness in a sample are correlated (Gotelli & Colwell, 2001), rarefaction techniques were utilized [using the EcoSim programme of Gotelli & Entsminger (2009)] to rarefy and compare fish species richness between regions (southern Persian Gulf and Gulf of Oman).Rather than arbitrarily using a single number of individuals, species richness was rarefied at 1000, 2000, 3000, 4000 and 5000 individuals.The programme randomly chooses 1000 individuals, for example, and then determines species richness.The procedure is repeated 1000 times to find the average expected species richness (±95% c.i.).Fish size-class distributions were also compared between regions using χ 2 tests on the proportional abundance of fishes in each L T class, based on the mean abundance of fishes in transects at each site, averaged across sites in each region.
Differences in multivariate fish community structure between the southern Persian Gulf and the Gulf of Oman, both in terms of species and guild composition, were compared graphically using ordination scatterplots from non-metric multidimensional scaling (NMS), with significance of differences tested using analyses of similarity (ANOSIM) (Clarke & Warwick, 2001).ANOSIM produces a test statistic, R, which assesses the null hypothesis that there are no differences among groups; an R-value >0•5 indicates strong differences among groups (Clarke , 2001).Fish species and guilds driving differences among groups were then identified using similarity percentage analysis (SIMPER), which quantifies the contribution of each species or guild to dissimilarity between groups (Clarke & Warwick, 2001).
A linear regression was used to determine whether fish species richness, abundance or biomass was associated with the amount of coral cover.In addition, correlations in the multivariate structure of fish and coral communities were tested with RELATE (Clarke & Warwick, 1994), using both fish species and guild structure v. coral species and morphology structure.RELATE is a non-parametric comparative (Mantel-type) test which tests the hypothesis of no relation between two multivariate groups, and produces a Spearman's rank correlation statistic (ρ) and a P -value based on a comparison with a randomized set of samples.

RESULTS
Nested ANOVAs showed significantly lower fish species abundance and biomass in the southern Persian Gulf compared with the Gulf of Oman (ANOVA, F 1,56 = 22•3 and 43•1, respectively, P < 0•01 each) and significant variation among sites within regions (ANOVA, F 6,56 = 5•6, 12•7 and 13•8, respectively, P < 0•001 each) (Table I).Rarefaction showed that Gulf of Oman communities held substantially higher species richness, when comparing the same number of individuals, to southern Persian Gulf communities.When sampling 4000 individuals in the southern Persian Gulf (the maximum number of individuals found for this region was 4134), the expected richness was 28•9 species (±0•1 95% c.i.).Sampling the same number of individuals in the Gulf of Oman, species richness was 61•6 species (±4•0 95% c.i.).
Ordination indicated substantial differences in the structure of fish communities between the southern Persian Gulf and the Gulf of Oman both in terms of species and guild composition (Fig. 2), and these differences were shown to be strong and significant in ANOSIM (species composition: R = 0•78, P < 0•001; guild composition: R = 0•74, P < 0•001).SIMPER indicated 82•1% dissimilarity in species composition between regions; species driving these differences are indicated in Table II.Of the 28 species that contributed to community dissimilarity, over half were more abundant in the Gulf of Oman than the southern Persian Gulf (Table II).The size structure of fish communities differed significantly between the southern Persian Gulf and the Gulf of Oman (Fig. 3; χ 2 = 85•8, d.f.= 7, P < 0•001).There was relatively higher abundance of fishes in the 50-100 mm L T class in the Persian Gulf, with abundance in all other L T classes higher in the Gulf of Oman.There were virtually no fishes (0•3%) >200 mm L T in the southern Persian Gulf, while there was an order of magnitude more fishes in >200 mm L T classes in the Gulf of Oman, with fish abundance showing a near exponential decline with each successive size-class here compared with the truncated distribution in the southern Persian Gulf (Fig. 3).
In terms of feeding guilds, there was 78•8% dissimilarity in guild structure between regions, with nearly half of this dissimilarity resulting from differences in the abundance of planktivores and grazers between regions (Table III).Nested ANOVAs indicated that only grazers were more abundant in the southern Persian Gulf (F 1,56 = 193•9, P < 0•001), while there were significantly more planktivores, piscivores, •6
Coral cover was comparable between the southern Persian Gulf and the Gulf of Oman (F 1,56 = 3•6, P > 0•05), but showed variation among sites within regions (F 6,56 = 8•1, P < 0•001; Fig. 5).Regression analyses showed that there was a significant relationship between fish abundance and total coral cover in transects across regions (r 2 = 0•08, d.f.= 62, P < 0•05), but there was no relationship between coral cover and fish biomass or fish species richness (r 2 = 0•02 and 0•01, respectively, d.f.= 62, P > 0•05 each).In terms of overall community structure, fish and coral communities were shown to be only weakly correlated with each other in RELATE in terms of both fish species and guild structure v. coral species and growth form structure (Table IV).

DISCUSSION
The southern Persian Gulf is characterized by extreme environmental conditions, with physical extremes not observed in tropical reefs anywhere else in the world (Sheppard et al., 1992;Coles, 2003;Riegl, 2003).Despite these conditions, diverse and abundant coral and coral-associated fish communities do develop in this area (Riegl, 1999;Burt et al., 2008), indicating that a variety of reef fauna can survive at or above the elevated sea temperatures predicted for much of the world in the coming decades (IPCC, 2007).This observation, however, is not without a note of caution: the results of the work here, and earlier work on benthic communities in the region (Sheppard, 1987;Sheppard & Sheppard, 1991;Coles, 2003), show that while a number of species can tolerate these extreme conditions, the communities are generally lower in diversity and abundance, and function differently, than those in more benign environments (Sheppard et al., 1992;Coles, 2003).
Although there was a significant relationship between fish abundance and total coral cover, the majority of fish assemblage parameters (i.e.biomass, richness and guild structure) had little relationship to coral cover, coral species structure or growth forms throughout regions.Such low concordance between demographic variables and coral assemblage structure shows that assemblages within both regions may be less tightly associated with the live coral structure than the reef structure itself (Graham et al., 2006).Indeed, there is still little agreement between researchers on the importance of live coral v. reef complexity in structuring reef fish communities (Graham et al., 2006;Feary et al., 2007;Pratchett et al., 2008).Particularly within the southern Persian Gulf, but also within the Gulf of Oman, substantial seasonal changes in coral cover can occur, mainly through overgrowth of turfing and fleshy macroalgae in spring months (Riegl, 1999;Burt et al., 2008), but also through warm-water bleaching (Riegl, 1999) and harmful algal bloom development (Richlen et al., 2010).Such seasonal variations in live coral cover may have resulted in fish assemblages becoming adapted to non-coral-dominated communities, where both habitat availability and feeding resources are more associated with seasonal changes in available benthic resources than with a stable coral reef platform (Riegl, 1999).
Despite little association between coral and fish community structure, there were broad differences in the abundance and biomass of fish communities between regions, which may be associated with either temperature or salinity effects (i.e.osmotic stress) on fish physiology (Hurst, 2007).Temperature and salinity are the primary variables regulating aquatic species bioenergetics (Kitchell et al., 1977), and fluctuations in ambient levels of either variable can substantially affect the rate and efficiency of physiological processes such as feeding, respiration, faecal egestion rates and ultimately growth (Kitchell et al., 1977;Hurst, 2007).Even under natural conditions, fishes will experience a range of temperature and salinity both spatially and temporally due to behavioural activities, and such variations may have substantial effects on the physiology of animals (Hurst, 2007).Within the southern Persian Gulf, with rapid temporal fluctuations between maximum and minimum temperature between seasons and amid relatively high background salinity levels (which fluctuate with season), the physiological optima for fishes may exist for limited periods during the year (i.e. between summer and winter).Therefore, key bioenergetic variables controlling physiological mechanisms may only be maximized for a short time frame during their annual cycle (Hurst, 2007;Figueira et al., 2009), and have a substantial effect on the development and structure of fish communities.Under predictions of increasing fluctuations in global oceanic temperature and salinities (IPCC, 2007), the structure of tropical reef fish communities may then undergo increasing periods of physiological stress (Munday et al., 2008;Nilsson et al., 2009) which may affect the demographic mechanisms important in structuring assemblages, and have flow-on effects to the abundance and biomass of communities (Munday et al., 2008).
High average sea-surface temperatures can have a substantial effect on plankton community composition; communities can become less productive within higher temperatures as longer, less-productive planktonic food chains are favoured (McKinnon et al., 2007).In this respect, levels of primary productivity differ between regions, with much higher productivity in the Gulf of Oman than Persian Gulf, due to upwelling of cold and nutrient-rich waters along the Omani coastline associated with the summer monsoonal conditions within the Arabian Sea (Schils & Wilson, 2006).In contrast, minor upwelling events combined with little riverine input have resulted in the Persian Gulf being a relatively nutrient-poor environment (Sheppard et al., 1992;Nezlin et al., 2007).Such divergence in primary productivity, associated with differences in physical variables controlling productivity between regions (Sarmiento et al., 2004;Brown et al., 2010), may underpin differences in the abundance and diversity of fish communities between regions (Brown et al., 2010).There is now increasing evidence that warming of oceanic waters can have a substantial effect on the distribution and abundance of ocean primary production (Richardson & Schoeman, 2004;Behrenfeld et al., 2006) and these changes have been linked to changes in higher trophic level organisms (Richardson & Schoeman, 2004;Beaugrand et al., 2008).Within the Persian Gulf, the physical extremes (i.e.high maximum summer temperatures) may have a substantial negative effect on the primary productivity within the Persian Gulf, which may then have flow-on effects to the abundance and diversity of fishes within this region (Richardson & Schoeman, 2004;Beaugrand et al., 2008).
There were broad differences in the size structure of assemblages between regions, indicative of a low average body size of fish assemblages within the southern Persian Gulf.Although both regions were dominated by relatively small-bodied fishes, the Gulf of Oman held significantly higher abundances of fishes >200 mm L T .Although there is little information on fishing at the sampling sites in either region, anecdotal evidence suggests that all sites surveyed are exposed to relatively low or moderate fishing activity.Therefore, the lower abundance of large-sized fishes in the Gulf of Oman was not likely to be due to fishing effects.Such differences in assemblage size structure may then be influenced by a complex interaction between both environmental and demographic factors (Fonseca & Cabral, 2007;Feary et al., 2009).In general, however, individuals tend to reach smaller maximum sizes and have shorter lives in physically stressful environments (Hurst, 2007;Munday et al., 2008), and recent evidence suggests that species within such environments may show reduced growth rates (Figueira et al., 2009).Reductions in species growth rate may have substantial effects on the demographic processes structuring their populations.In particular, there will be an increase in time to sexual maturity (Booth, 1995), reducing lifetime gamete production and potentially both local and regional replenishment.In addition, reductions in growth rate may also indirectly affect species mortality, increasing the length individuals remain in small, predator-vulnerable sizes thereby enhancing their potential for predation risk (Holbrook & Schmitt, 2002).While such theoretical changes in fish communities within physically stressful environments are yet to be rigorously tested (Munday et al., 2008;Wilson et al., 2010), communities utilizing such environments (such as within the southern Persian Gulf) provide a unique opportunity to examine such predictions and will be important in understanding the potential repercussions of climate change on the ecological structure of marine communities.
This study has shown that the southern Persian Gulf has much lower fish species richness than the adjacent Gulf of Oman.This finding is in corroboration with similar studies arguing that this fauna represents a biogeographic subset of those in the Indian Ocean that have re-colonized the area following flooding after the last glacial period (Sheppard & Sheppard, 1991;Coles, 2003).Such low richness, however, is likely also due to the physical extremes within the Persian Gulf (e.g.variation in temperature and salinity) for which few Indian Ocean species have been able to adapt and tolerate (Coles & Tarr, 1990;Randall, 1995).In corroboration, even within the Persian Gulf patterns in species richness are associated with changes in physical extremes.For example, within the deeper waters of the northern part of the Persian Gulf (along the Iranian coastline), where variations in temperature and salinity are less extreme, communities are noted as being richer in species than waters in the southern region (Price et al., 1993).In parallel, in the south-western basin, highest © 2010 The Authors Journal of Fish Biology © 2010 The Fisheries Society of the British Isles, Journal of Fish Biology 2010Biology , 77, 1931Biology -1947 species richness has been observed near the deep-water offshore islands of Saudi Arabia, with diversity declining north and east of this area, probably as a result of more extreme water salinity (Price et al., 1993).
Guild structure changed considerably from the southern Persian Gulf to the Gulf of Oman.This was predominantly due to lower abundances of most trophic guilds within the southern Persian Gulf compared with the Gulf of Oman, and a higher abundance of herbivorous grazers.It can be predicted that differences in guild structure may be associated with seasonal changes in benthic habitat, with the southern Persian Gulf holding relatively high cover of seasonally occurring turfing and fleshy macroalgae (Riegl, 1999;Burt et al., 2008).In this respect, the much higher numerical abundance of corallivores within the Gulf of Oman will be due to more widespread coral cover throughout this coastline (Sheppard et al., 1992;Coles, 1997Coles, , 2003)).Reduced abundances of piscivores in the southern Persian Gulf are more likely to be due to reductions in the average abundance of fish assemblages within this region, as piscivore abundance will usually mirror prey abundance (Stewart & Jones, 2001).Substantial differences in the abundance of planktivores between regions were also found, with substantially smaller abundances within the southern Persian Gulf.As the majority of planktivores within the eastern Arabian Peninsula are smaller bodied fishes (i.e.Pomacentridae and Caesionidae), it can be predicted that this trophic group may be more negatively affected by extremes in environmental conditions within the southern Persian Gulf (i.e.high temperature variations) than larger species, due to the metabolic demands of smaller size fishes (Munday & Jones, 1998).
Understanding whether the patterns reported in this work are applicable to reefs globally will be important in determining the wider implications of this research.Between the Gulf of Aden and southern Red Sea regions, similarities in the change in physical variables to that found between the southern Persian Gulf and Gulf of Oman are present (Roberts et al., 1992).Both temperature and salinity show an increasing gradient moving from the Gulf of Aden to the Red Sea, with average summer temperatures of 28 and 30 • C and summer salinities of 36•5 and 42•0, respectively (Roberts et al., 1992;Sheppard et al., 1992).There is also a change in productivity between regions, with inflowing waters from the Gulf of Aden being richer in nutrients than the oligotrophic waters of the Red Sea, with an abrupt drop in nutrient concentration north of 19 • N (Roberts et al., 1992).Such changes in physical variables have resulted in a zoogeographic boundary for species distribution, with substantial reductions in the abundance and diversity of reef fish communities when comparing the Gulf of Aden and the Red Sea (Roberts et al., 1992;Kemp, 1998).Although low levels of available coral reef habitat in the southern Red Sea may have a substantial effect on reef fish abundance and diversity (Roberts et al., 1992;Kemp, 1998), such differences in reef fish community structure may also be associated with the substantial changes in physical variables between regions (Roberts et al., 1992).
There were substantial differences in the structure and composition of reef fish communities in the southern Persian Gulf and the Gulf of Oman, despite these being biogeographically connected, indicating the importance of environmental extremes in structuring fish assemblages.Although the possible role of high water temperature in determining community composition and abundance has been emphasized, extremes in salinity levels, along with relatively low primary productivity will also act synergistically to have substantial effects on the structure and assembly of southern Persian Gulf communities.Although there is still little research investigating the © 2010 The Authors Journal of Fish Biology © 2010 The Fisheries Society of the British Isles, Journal of Fish Biology 2010Biology , 77, 1931Biology -1947 role of fluctuations in ocean processes in structuring reef-associated fish communities (Roessig et al., 2004;Harley et al., 2006;Munday et al., 2008), the results of this work suggest that measurable differences in major oceanographic variables dominating regions may have substantial flow-on effects to the composition and structure of the associated reef fish community.
Logistic support for this project was provided by the Environment Agency of Abu Dhabi, Nakheel PJSC and the Fujairah Municipality, and their support is appreciated.A portion of this study was part of the Nakheel -UN University International Network on Water, Environment and Health (UNU-INWEH) research programme undertaken to understand the ecological dynamics of ecosystems in the vicinity of Nakheel marine projects.

Fig. 3 .
Fig. 3. Relative abundance of fish in each total length (L T ) class in the southern Persian Gulf ( ) and the Gulf of Oman ( ).
Journal of Fish Biology © 2010 The Fisheries Society of the British Isles, Journal of Fish Biology 2010, 77, 1931-1947F I S H C O M M U N I T I E S O N T H E W O R L D ' S WA R M E S T R E E F S 1939Table III.Guild differences in fish community structure between the southern Persian Gulf and the Gulf of Oman.Mean abundance per 150 m 2 in each region and individual and cumulative contribution (δ %) to dissimilarity from SIMPER are Fig. 4. Mean ± s.e.abundance of fishes in feeding guilds at sites within the southern Persian Gulf ( ) v. the Gulf of Oman ( ).Sites, in order of appearance (left to right), are Ras Ghanada, Sadiyaat, Ras Hasyan and Saih Al-Shaib (southern Persian Gulf), and Dibba, Al-Aqah, Al-Ghattan and Al-Jazira (Gulf of Oman).

Table I .
Mean ± s.e.species richness and abundance at each site within the Persian Gulf and Gulf of Oman.Different superscript lower case letters indicate significant differences between sites within each region based on Tukey's tests Richness , 2010, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1095-8649.2010.02777.xby Nova Southeastern University, Wiley Online Library on [19/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License The Authors Journal of Fish Biology © 2010 The Fisheries Society of the British Isles, Journal of FishBiology 2010Biology  , 77, 1931Biology  -194710958649

Table IV .
Spearman's rank correlation (ρ)between multivariate fish communities, in terms of species and guild structure, and coral communities, in terms of species and growth forms, based on RELATE analyses (P < 0•001) , 2010, 8, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1095-8649.2010.02777.xby Nova Southeastern University, Wiley Online Library on [19/04/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License © 2010 The Authors Journal of Fish Biology © 2010 The Fisheries Society of the British Isles, Journal of FishBiology 2010Biology  , 77, 1931Biology  -1947